The Agent Skills Directory

[SAFE]: The skill provides structured guidance and core principles for designing effective agent skills. The instructions are benign, professional, and focus on context efficiency.\n- [EXTERNAL_DOWNLOADS]: The skill includes references to trusted external resources, including official Anthropic support pages (support.claude.com), documentation (resources.anthropic.com), and GitHub repositories (github.com/anthropics). These are documented as safe references to well-known service providers.\n- [NO_CODE]: The instructions reference utility scripts (init_skill.py and package_skill.py) for project management tasks. These scripts are vendor-specific resources for NeverSight but were not included in the analyzed file set. The provided files consist entirely of non-executable markdown and text.\n- [PROMPT_INJECTION]: Analysis of the instructional content confirmed no attempts to bypass safety filters, extract system prompts, or override agent constraints. The skill uses standard instructional language for its intended purpose.

skill-creator